A revision of the chelonian genus Bothremys (Pleurodira: Pelomedusidae)

UNIVERSITY OF

ILLINOIS LIBRARY

ATURBANA-CHAMPA1GN

GEaOGY

The person charging this material is re-
sponsible for its return to the library from
which it was withdrawn on or before the
Latest Date stamped below.

Theft, mutilation, and underlining of books
are reasons for disciplinary action and may
result in dismissal from the University.

UNIVERSITY OF ..MHOIS UBRARY AT URBANA-CHAMPAIGN

OCT ^3
0(^19

I L161 — O-1096

/

7 FIELDIANA • GEOLOGY

Published by
FIELD MUSEUM OF NATURAL HISTORY

Volume 16 March 21, 1968 No. 7

A Revision of the Chelonian Genus Bothremys
(Pleurodira: Pelomedusidae)

Eugene S. Gaffney

Preceptor, Department of Geology, Columbia University
AND

Rainer Zangerl
Chief Curator, Department of Geology, Field Museum

ACKNOWLEDGEMENTS

The senior author in this study was supported by a K.P. Schmidt
Museum Fellowship and a National Science Foundation Summer
Fellowship for Graduate Teaching Assistants. We would like to ex-
press our sincere thanks to the following persons for their kind-
ness and generosity: Dr. Donald Baird of Princeton University read
the manuscript and helped considerably throughout the investiga-
tion. His advice and encouragement has greatly increased the value
of this study. Dr. Samuel B. McDowell of Rutgers University read
the manuscript and made a number of helpful suggestions. Dr. Hor-
ace Richards of the Academy of Natural Sciences lent us the Mer-
chantville specimen of Bothremys. All of the photographs (except
Figs. 2, 3, 18, 22D,E) were taken with equipment lent by Mr. Har-
old Rollins of the Department of Fossil Invertebrates, American
Museum of Natural History. We would also like to thank Dr. E. H.
Colbert for advice and allowing one of us to use the American Mu-
seum of Natural History's facilities. The line drawings were made
by Mrs. Maidi Wiebe Leibhardt and the senior author.

INTRODUCTION

In the fall of 1945 the late CM. Barber with the help of Karl P.
Schmidt, then Chief Curator of Zoology, and I instigated a program

Library of Congress Catalog Card Number: 68-2X236
No. 1041 H»,WS.t,f5Jlli«

JUL 11 «68
GEOLOGY UBRAK* ^

194 FIELDIANA: GEOLOGY, VOLUME 16

of collecting vertebrates in the Mooreville Chalk of Alabama. By
1948 most localities had been thoroughly searched and most of them
had been revisited a number of times, but Mr. Barber proposed one
last trip in the fall of 1949 for which he and I received a grant from
the Geological Society of America (No. 661-55). This is herewith
gratefully acknowledged.

It was on this occasion that Mr. Barber, accompanied by Mr. J.
A. Robbins, collected a most significant specimen, a virtually com-
plete shell of the pelomedusine Podocnemis alabamae associated with
a partial skull and a mandible (FMNH PR 247). This mandible
clearly showed that the material heretofore assigned to Podocnemis
alabamae properly belongs to the genus Bothremys Leidy.

While visiting collections of East Coast Institutions in connection
with my study of the protostegid and toxochelyid turtles of this
fauna, I discovered in the Peabody collection, Yale University, a tray
full of turtle shell fragments from the Niobrara Chalk of Kansas, and
recognized them as belonging to a pelomedusine closely allied to the
Alabama and Arkansas forms. The specimen has since been piece-
fitted, and is described below.

In casual conversation with Dr. Horace Richards of the Phila-
delphia Academy of Sciences, I suggested that the turtles of the New
Jersey Greensand probably occur often in whole shells, and that
the fragmentary condition of most of the material from New Jersey
is due to casual collecting. Dr. Richards instructed interested young
collectors, Albert Jehle and Henry B. Roberts, in the proper collec-
tion of fossil material in the Greensand pits and this brought forth
striking results. Within a few months, the bony plates of a large tur-
tle arrived at Field Museum and these could easily be piece-fitted to
form a nearly complete shell. This find also belongs to a pelomedu-
sine closely related to the Alabama and Arkansas forms. The revi-
sions made necessary by these materials are in the present paper.

Rainer Zangerl

EMENDED DIAGNOSIS

Order Chelonia

Suborder Pleurodira

Family Pelomedusidae

Genus Bothremys Leidy, 1865

Fig. 1. Correlation chart.

N iobrar a F m
B. barberi s u bsp.C

'19 52/, Dor f / 1 9 5 7/
and Cooke. etal/1943/

— CO

CO Q.

5 »
x>

c =>
S «
— o

C "to

* ?

o <o

*- -«a

to

f t er Cobban eta 1.
tephanson eta 1. /1 942

Moorevi lie Tongue of

S e 1 ma Fm
B barber i subsp. B

T

1

Hornerstown Fm/lower/

Merchant vi lie F m
B. barberi s ubsp. A

•

o

i

E
>

<D

c

<D
O
O

a.

a>

a>
o
o

2

a>

a>
o
o

GO

o

a>
o
o

UJ

c
a>
o
o
a>

co

a.

<_>

10

CD

to

s

c
co

c
to
Q.

E
to
O

c
to

c
o

c/5

c

(0

o

CO

c
o
o

co

c
o

1-

A j e 1 } J 9 i

snoa oejajQ

195

196

FIELDIANA: GEOLOGY, VOLUME 16

Fig. 2. Bothremys barberi, spp. B (FMNH PR 247), Selma Formation, Ala-
bama, dorsal view of carapace. Shell is flattened and distorted, the width is greater
than it was in life. N - neural, NUC - nuchal, PE - peripheral, COST - costal,
PG - pygal, SPG - suprapygal.

Type species: Bothremys cooki

Skull wide and flat, orbits facing dorsad. Palate dominated by-
large pits in the center of wide triturating surfaces. Pits formed
by jugals widely exposed on triturating surfaces. Well developed
vomer separating internal nares. Processus pterygoideus externus as
in living pelomedusids. Quadrate closed behind stapedial opening.

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 197

Fig. 3. Bothremys barberi, spp. B (FMNH PR 247), Selma Formation, Ala-
bama, ventral view of plastron. E - epiplastron, ENT - entoplastron, HYO -hyo-
plastron, HYPO - hypoplastron, ME - mesoplastron, X - xiphiplastron. Sagittal
length is 48.5 cm.

Probable contact of quadrate with basisphenoid. Processus articu-
laris on same vertical plane with condylus occipitalis. No enlarged
carotid opening.

Known appendicular elements similar to Podocnemis expansa.
Carapace oval, moderately arched. Nuchal hexagonal, broader pos-
teriorly, anterior edge notched. Neurals vary from six to seven.

198 FIELDIANA: GEOLOGY, VOLUME 16

Costals 2-4 tend to be much shorter in relation to their width as
in Taphrosphys, but as opposed to Podocnemis. Eleven peripherals,
free edges mostly acute but varying to obtuse in large individuals.
No cervical scute, pleural scales extending onto peripherals. Anterior
lobe of plastron short, broad, and rounded. Posterior lobe long,
tapering posteriorly. Xiphiplastral notch broad and U-shaped, xi-
phiplastra end in rounded projections. Large, quadrilateral ento-
plastron. Mesoplastra small and well onto bridge. Pubic scar short
and wide, ischiac scar triangular and away from xiphiplastral edge.
Intergular V-shaped and separating small gulars and part of hu-
merals.

The species cannot be diagnosed at this time.

DISCUSSION OF SPECIES

Bothremys cooki Leidy, 1865

A skull and mandible (Figs. 13, 14, 15, 16, 22) are the only known
specimens of this species. It was found in 1862, described by Leidy
in 1865 (pp. 110-113), and re-examined by Hay in 1908 (pp. 103-
104). The original locality data (Leidy, 1865, p. 110) reads: "the
Green-sand near Barnsboro, Gloucester County, New Jersey." The
skull's age was presumed to be Late Cretaceous, but subsequent
revisions of New Jersey stratigraphy show that the age could be
either Late Cretaceous or Lower Tertiary (see Miller, 1956; Olsson,
1963; and Baird, 1964). However, the most prolific vertebrate hori-
zons in this area are the lowermost Hornerstown Formation and the
Navesink Formation, both of which are Maestrichtian.

Even though the jaws of B. cooki and B. barberi (see below) are
very similar, the authors think it best to retain the species distinc-
tion until associated shell material becomes available for B. cooki.

Bothremys barberi (Schmidt)

Podocnemis barberi Schmidt, 1940

Podocnemis alabamae Zangerl, 1948
The discovery of a lower jaw (Fig. 22) almost identical to Both-
remys cooki associated with a nearly complete carapace and plastron
(Figs. 2, 3) of "Podocnemis alabamae" has made it necessary to change
the generic assignment of "P. alabamae." The differences between
"P. barberi" and "P. alabamae" 'seem trivial in the light of material
from New Jersey and Kansas; we thus believe that these two forms

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS

199

Fig. 4. Bothremys barberi, spp. A (ANSP 15902), Merchantville Formation,
New Jersey, dorsal view of carapace.

are conspecific. "P. alabamae" becomes a synonym of "P. barberi,"
as the oldest available name for this species.

The generic re-assignment is based on a weathered but almost en-
tirely complete specimen (Figs. 2, 3) collected by Mr. CM. Barber
in the Mooreville Tongue of the Selma Formation, one mile east of
Harrell, Dallas Co., Alabama (FMNH PR 247). The shell is read-
ily identified as "Podocnemis alabamae." The associated skull ma-

200

FIELDIANA: GEOLOGY, VOLUME 16

Fig. 5. Bothremys barberi, spp. A (ANSP 15902), Merchantville Formation,
New Jersey, ventral view of carapace.

terial consists of a complete lower jaw, water-worn around the edge,
plus portions of the brain case and right otic capsule described below.
Comparison of B. barberi's jaw with the type of Bothremys cooki re-
veals no important differences.

As can be seen from the correlation chart (Fig. 1) there are a
number of specimens of B. barberi occurring in formations of about
the same stratigraphic horizon from New Jersey to Kansas. This
suggests the possibility of subspecies along the Cretaceous coastline.
Such an arrangement would be feasible in view of subspecific dif-
ferentiation of modern turtles along the Gulf Coastal Plain (e. g.,

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 201

Graptemys, where virtually every major river has its distinct sub-
species) . In the case of the Cretaceous coastline, it seems not too un-
reasonable to distinguish Atlantic, Gulf, and epicontinental coastal
provinces, especially since the Gulf province was deeply indented
by the Mississippi embayment. In view of what is known of sub-
speciation in modern turtles, it is likely that if large enough samples
were available, these three groups could be defined statistically.
However, since two of the samples (New Jersey and Kansas) consist
only of one specimen each, it is impossible to differentiate between
variation on a population level and individual variation. As a com-
promise between theoretical considerations and the practical limits
of the material, we propose to designate these subspecies by letters
but not name them formally.

Subspecies A: The Atlantic Coastal Province

A nearly complete shell from New Jersey (ANSP 15902, Figs.
4-8) extends the geographic range of the species to the East Coast.
The specimen was found in the Merchantville Formation at Graham
Brick Yards, Maple Shade, New Jersey, in 1945, by Albert Jehle and
Henry B. Roberts (Dr. Horace Richards, personal communication).
The Merchantville Formation consists of "black, glauconitic, mica-
ceous clay with occasional sand layers" (Dorf, 1957, p. 192) and is
believed to be lower Campanian in age (ibid.) or upper Santonian
(Stephenson et al., 1942). In addition to this specimen, the known
vertebrate fauna of the Merchantville (Miller, 1955, p. 904) includes
fish teeth, mosasaur fragments, a possible Trionyx, and crocodile
remains. The only other record of a specimen possibly assignable
to this genus from New Jersey is a Podocnemis sp. recorded by Miller
(1955, p. 907). Upon examination the material (ANSP 9370) is
seen to consist of fragments which can be determined as being pelo-
medusid, but a generic identification is not possible.

The shell has not been subjected to distortion or compression.
However, some sutures have been obscured by a limonitic cement.
The plastron is nearly complete and well preserved. The left ilium
and part of the right are preserved as well as a small portion of the
pubis.

The ventral plates of the bridge peripherals (Fig. 6) in the Gulf
shells are nearly as long as or slightly less than the dorsal plates.
In the Merchantville turtle the ventral plates are about half as
long as the dorsals. The hinder peripherals and pygal agree in out-
line with the southern forms, but a thick, ventral ridge, parallel to

##»*% POST.

VENTRAL

Fig. 6. Bothremys barberi, spp. A (ANSP 15902), Merchantville Formation,
New Jersey, ventral view of right peripherals 3-6, sections made at points indi-
cated by dashes.

202

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 203

Fig. 7. Bothremys barberi, spp. A (ANSP 15902), Merchantville Formation,
New Jersey, ventral view of plastron.

the edge of the carapace is developed on peripherals ten, eleven, and
the pygal. This ridge is absent in the Gulf turtles. The sulci are de-
terminable only in a few places and agree with previously described
specimens of B. barberi (Podocnemis alabamae and P. barberi.) A
small anomalous bone is present on the right bridge between meso-
plastron and hyoplastron. The diagnostic features indicating the

204

FIELDIANA: GEOLOGY, VOLUME 16

Fig. 8. Bcthremys barberi, spp. A (ANSP 15902), Merchantville Formation,
New Jersey, dorsal view of plastron.

generic identity of the Merchantville specimen with Bothremys may
be summarized as :

1) costals 2-4 tend to be much shorter in relation to their
width than in related genera;

2) anterior lobe of plastron short, broad and rounded;

3) posterior lobe long, tapering posteriorly;

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS

205

i^«Mv # IP
h : tfk, rf . . r 7; It

£&

fl--f?

*ZKi

Fig. 9. Bothremys barberi, spp. C (Peabody Museum 3608), Niobrara Forma-
tion, Kansas, dorsal view of carapace.

4) xiphiplastral notch broad and U-shaped, xiphiplastra end

in rounded projections;

5) large, quadrilateral entoplastron ;

6) pubic scar short and wide, ischiac scar triangular and sep-
arated from xiphiplastral edge.

Subspecies B: The Gulf Coastal Province

The forms described by Schmidt (1940) and Zangerl (1948) as
"Podocnemis barberi" and "P. alabamae" make up this subspecies.

206

FIELDIANA: GEOLOGY, VOLUME 16

Fig. 10. Bothremys barberi, spp. C (Peabody Museum 3608), Niobrara Forma-
tion, Kansas, ventral view of carapace.

The new shell FMNH PR 247 (Figs. 2, 3) agrees with the previously
described material from the Selma Formation.

Subspecies C: The Epicontinental Province

A shell in the Peabody Museum belongs to Bothremys barberi
(Figs. 9-12). The specimen (Peabody Museum 3608) was collected in
the Niobrara Formation of Wallace Co., Kansas, on June 14, 1876,
by B. F. Mudge and party. It consists of a nearly complete plastron,
lacking portions of the right bridge and portions of both xiphiplastra,
as well as about 60 per cent of the carapace. The anterior end of the
carapace is complete and includes the first two peripherals. Six
neurals and the proximal portions of the first six costals are also
preserved. The shell is crushed, causing the ridge area to lie flat in
the plane of the plastron. The dorsal surface of the carapace is ex-

GAFFNEY AND ZANGERL: THE GENUS ROTHREMYS 207

Fig. 11. Rothremys barberi, spp. C (Peabody Museum 3608), Niobrara For-
mation, Kansas, ventral view of plastron.

tensively damaged by deep pits, apparently of parasitic origin. Other
evidences of parasitism are linear indentations on both plastron and
carapace.

There are three main morphologic distinctions separating this
specimen from the New Jersey and Gulf forms:

1) the nuchal is longer and narrower (in restoration the nu-
chal has been widened somewhat, so this is not as apparent
in the figure as it should be) ;

2) the mesoplastra are equidimensional while in the other
forms, width exceeds the length;

3) the epiplastral symphysis is relatively longer.

208

FIELDIANA: GEOLOGY, VOLUME 16

Fig. 12. Bothremys barberi, spp. C (Peabody Museum 3608), Niobrara For-
mation, Kansas, dorsal view of plastron.

The scale pattern agrees with the other specimens except for the
anterior marginals which are nearly equidimensional, not wider than
long as in the Gulf shells. An anomalous scale is present between
the first vertebral and the first marginals. (Fig. 9) .

A Miocene occurrence of Bothremys

A pelomedusid described by Collins and Lynn (1936, p. 155, plate
I) as Taphrosphys miocenica seems to belong to the genus Bothremys
instead. The type and only known specimen is the anterior portion
of a plastron (USNM 13784) collected in 1934 from zone 10 of the

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 209

Calvert Formation, about one-fourth mile south of Camp Roosevelt,
Maryland. Although the type cannot be found at present (Dr.
Nicholas Hotton III, personal communication), the published fig-
ures and description are adequate to show that the specimen belongs
to Bothremys. In Taphrosphys the large intergular separates the gulars,
humerals, and part of the pectorals. It is not possible to determine
the species since the type consists of such a small amount of material.

It should be emphasized that this specimen is the only North
American pleurodire definitely known from Tertiary rocks. The
genus Taphrosphys has been reported from the Lower Tertiary of
New Jersey (Zangerl, 1948; Romer, 1956), but the stratigraphic data
are questionable and specimens with good locality data are
Cretaceous.

THE SKULL OF BOTHREMYS
INTRODUCTION

There are no pleurodires known from pre-Cretaceous sediments.
Besides Bothremys the only Cretaceous pelurodiran skull materials
known to the authors are fragments of Taphrosphys (being described
by Gaffney) and undescribed Brazilian specimens being worked on
by Price. The authors feel that since Bothremys has a well-pre-
served skull and since it is one of the oldest pleurodires, a rigorous
description is necessary to provide a basis for comparison.

This description follows the anatomical nomenclature for turtle
skulls standardized by Parsons and Williams (1961).

The material described here consists of a nearly complete skull
with mandible (the type of B. cooki) and a partial skull with com-
plete mandible (B. barberi FMNH PR 247). Where elements present
in both skulls are described, B. cooki will be described first. The
following recent material, belonging to the Department of Living
Reptiles and Amphibians, American Museum of Natural History,
served as comparative material in this study. The authors would
like to thank Drs. Bogert and Zweifel for supplying these specimens.

F. Pelomedusidae:

Pelomedusa subrufa, AMNH 63581 ; Pelusios gabonensis, AMNH
10056; Pelusios subniger, AMNH 71188 and 10062; Podocnemis ex-
pansa, AMNH 58098 and 58194; Podocnemis madagascariensis,
AMNH 63579; Podocnemis unifilis, AMNH 58195 and 58200.

210 FIELDIANA: GEOLOGY, VOLUME 16

F. Chelyidae:

Chelodina longicollis, AMNH 76569; Chelus fimbriata, AMNH
1117 (Vertebrate Paleontology Comparative Anatomy Collection);
Platemys platycephala, AMNH 75101; Emydura sp., AMNH 72418;
Phrynops nasuta, AMNH 58123 and 58110.

GROSS ASPECT
B. cooki

The type skull of Bothremys cooki (Figs. 13-16) lacks the postero-
lateral portions including occiput and quadrate, but the remainder
is well preserved and undistorted.

In contrast to living pelomedusids, the skull of Bothremys is very
wide, being about as wide as it is long (excluding crista supraoccipi-
talis) . This width is due principally to the presence of large paired
pits in the triturating surfaces. The short face is comparable to liv-
ing pelomedusids, but the orbits are farther apart and face more
dorsad in Bothremys.

The bones roofing the fossa temporalis are missing. There is no
indication of emargination, the posterior limits of the skull roof
being represented by broken edges. The orbits are nearly circular
and face antero-dorsolaterad.

The dorsal surface of the skull is horizontal posteriorly, slop-
ing somewhat anteriorly. There is a slope of about 25 degrees from
between the orbits to the apertura narium externa. The apertura
is preserved best on the right side but still lacks the medial-most parts.
The opening has the shape of a figure eight lying on its side, similar
to Podocnemis expansa but lower and wider. There is no conclusive
evidence against the possibility that a thin septum was present medi-
ally. The anterolateral surface of the skull is broad and convex
outward.

The triturating surfaces of the palate are narrow anteriorly and
expand posteriorly and laterally to accommodate the huge pits pres-
ent between the lingual and tomial margins. The area between the
lingual ridges consists of a shallow median trough containing the
paired aperturae narium internae and the foramina praepalatina.

Processus pterygoidei externi are present on the pterygoids, but
since they are broken laterally their complete extent is not known.
The margins of the fenestra subtemporalis as preserved are roughly
comparable to those of Podocnemis expansa. The processus articu-
laris of the quadrate and condylus occipitalis are not preserved.

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 211

B. barberi

The processus articularis is placed farther posteriorly than in the
living pleurodires. The cavum tympani has a closed posterior margin
and is well developed. The incisura columellae auris is a highly ossi-
fied canal. The fenestra postotica cannot be determined.

No indication of sulci can be seen on the surface of either skull.
The dorsal surface of the parietals in B. cooki has a series of fine
branching grooves, many ending in small foramina (Fig. 14B). This
sort of surface may also be seen in Podocnemis.

PARIETAL
B. cooki

The paired parietals (Figs. 13, 14, 18) cover the median part of
the posterior skull roof. They meet the frontals anteriorly in a trans-
verse suture and the postorbitals more laterally as in the living pelo-
medusids. The lateral and posterior portions are represented by
broken edges; no indications of emargination are preserved.

The vertical component of the parietal (Figs. 13, 14, 16, 17) con-
sists of two pairs of plates, one pair on each side, which join dorsally
to form one pair meeting the horizontal component of the parietal.
Anteriorly the medial plate (processus inferior parietalis) is sutured
to the palatine near the posterior edge of that bone and borders the
relatively reduced orbito-cranial passage (no term available in Par-
sons and Williams, 1961). Farther posteriorly the parietal meets the
crista pterygoidea in a long, horizontal suture that ends posteriorly at
the foramen nervi trigemini. This foramen is bordered dorsally by
the parietal, but this bone contributes to less than one-fourth of the
circumference of the foramen. The most posterior suture is with the
supraoccipital.

The more lateral process of the parietal (Figs. 16, 17) diverges
from the processus inferior parietalis antero-ventrally and meets a
flange from the postorbital antero-laterally. It probably had a con-
tact with the pterygoid, but this area is now missing.

The ventral region of the parietal in living pelomedusids con-
sists principally of only one pair of vertical plates. The lateral plate,
which in Bothremys is formed by the parietal, consists of the postor-
bital, palatine, pterygoid, and sometimes parietal in the living genera.

Fig. 13. Bothremys cooki (type skull), ?Hornerstown Formation, New Jersey.
A, postero-lateral view (see Fig. 14A). B, dorsal view (see Fig. 14B). ANE -
apertura narium externa, BS - basisphenoid, F - frontal, FNT -foramen nervi trige-
mini, J - jugal, M - maxilla, P - parietal, PL - palatine, PO - postorbital, PPE —
processus pterygoideus externus, PR - prootic, PRF - prefrontal, PT - pterygoid,
SOC - supraoccipital.

212

B

,.«&»■

Fig. 14. Bothremys cooki (type skull), ?Hornerstown Formation, New Jersey.
A, postero-lateral view. B, dorsal view. C, anterior view. Scale - 1 cm.

213

214 FIELDIANA: GEOLOGY, VOLUME 16

The medial portions of the parietals are preserved in B. barberi
and agree with B. cooki except for being relatively thicker and more
massive.

FRONTAL

B. cooki and B. barberi

The frontal (Figs. 13, 14) is a quadrilateral element meeting,
from front to back, the prefrontal, orbits, postorbital, and parietal.
On the ventral surface a triangular ridge defines the sulcus olfacto-
rius (Fig. 17). The frontal has about half the length of the pre-
frontal or parietal, being relatively smaller, but otherwise quite sim-
ilar, to the living pelomedusid genera. There is no longitudinal
groove on the dorsal surface as in South American Podocnemis, with
the exception of P. dumeriliana.

PREFRONTAL

B. cooki

The prefrontal (Figs. 13, 14) is a long, somewhat irregular ele-
ment, sutured posteriorly to the frontal. It forms the antero-medial
margin of the orbit and contributes to about one-third of its circum-
ference. Only the right prefrontal is preserved anteriorly, and the
limits of the bone in this area are somewhat ambiguous. As we inter-
pret it, the prefrontal meets the maxilla in an irregular suture be-
ginning at the dorsal part of the apertura narium externa and run-
ning postero-laterally to the ventral edge of the orbit.

On the ventral surface the sulcus olfactorius continues for a short
distance on the prefrontal.

Bothremys has a short process descending ventrally along the
inner margin of the anterior orbital rim which meets the palatine
ventrally. In living pelomedusids this process no longer reaches
the palatine.

The orbito-cranial passage is very large in living pelomedusids
but is restricted in Bothremys because the skull roof, and consequent-
ly the sulcus olfactorius and processus inferior pterygoideus, are very
low. Also, the vomer rises above the level of the palate, further
restricting the fissura ethmoidalis.

The foramen orbito-nasale is not distinct from the fossa nasalis,
that is, the prefrontal does not descend to meet the vomer in the
midline. This condition is characteristic of all pleurodires.

GAFFNEY AND ZANGERL: THE GENUS BOTH REM YS 215

POSTORBITAL

B. cooki

The postorbital (Figs. 13, 14) consists of two plates at right an-
gles to each other. The dorsal plate meets the jugal antero-laterally,
the orbit antero-medially, the frontal medially, the parietal postero-
medially, and has a broken edge postero-laterally. The original
posterior extent of the postorbital is not known. Ventrally the
postorbital consists of a vertical plate, diverging anteriorly about
45 degrees from the skull midline. This plate has a long union with
the jugal antero-laterally, reaches the pterygoid postero-ventrally
and meets the lateral vertical process of the parietal posteriorly.
The parietal union is not preserved for half its ventral length.

The postorbital of living pelomedusids differs from Bothremys
principally in being less developed ventrally and contributing less
to the anterior wall of the temporal fossa.

JUGAL

B. cooki

The jugal has a fairly limited exposure dorsally, but its ventral
surface forms the pit so characteristic of this animal (Figs. 15, 16, 20) .
If removed from the surrounding bones, the jugal is somewhat cone-
shaped. Anteriorly the jugal meets the maxilla in an extensive and
complex pattern.

A well developed buttress extends antero-laterally from the cone-
shaped area of the jugal to the medial surface of the maxilla. The
posterior corner of this process abuts against the processus pterygo-
ideus externus of the pterygoid. Ventrally a thin sheet of palatine
covered the jugal-pterygoid suture, but is now broken off. Above
this suture the entire] posterior edge of the jugal lies against the
postorbital. Medial to the pterygoid contact there is a short suture
with the palatine. In front of this the maxillary suture begins.

The dorsal plate of the jugal lies between the maxilla and post-
orbital, and borders the orbit. The posterior-most edge is broken.

In living pelomedusids, the area homologous to the pits in
Bothremys faces posteriorly into the subtemporal cavity and does
not enter onto the triturating surface. This condition also prevails
in amphichelydians, and, in fact, in most reptiles. A complex (but
unknown) series of evolutionary events must have taken place to
make this area available as a triturating surface.

216

FIELDIANA: GEOLOGY, VOLUME 16

FPR

PM

V

mm

M

FPP->

PL

PT

PPE

BS

Fig. 15. Bothremys cooki (type skull), ?Hornerstown Formation, New Jersey,
ventral view of skull. Abbreviations also for Figures 16 and 17. BS - basisphenoid,
CL - cavum labyrinthicum, FNT -foramen nervi trigemini, FPP - foramen palati-
num posterius, FPR - foramen praepalatinum, ICA - incisura columellae auris, J -
jugal, LP - lateral process of parietal (see text under Parietal), M - maxilla, PIP -
processus inferior parietalis, PL - palatine, PM - premaxilla, PPE - processus ptery-
goideus externus, PT - pterygoid, SF - sulcus olfactorius, SO - supraoccipital, V -
vomer. Dotted areas are restored.

PREMAXILLA

B. cooki

Most of the right premaxilla is lost and the antero-medial part
of the left premaxilla is broken. The premaxilla (Fig. 15) is rough-
ly triangular in shape, being widest anteriorly, and forms the floor
of the apertura narium externa. Laterally it is sutured to the maxilla
for most of its length, but there is a short meeting with the vomer

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS

217

at the posterior end of the premaxilla. Just slightly anterior to the
vomerine suture and close to the midline lies the foramen praepal-
atinum. The postero-medial edge of the foramen is missing, and it
is not known whether it was completed by the premaxilla, maxilla
or vomer.

Fig. 16. Bothremys cooki (type skull), ?Hornerstown Formation, New Jersey,
ventral view of skull with basisphenoid and most of pterygoids removed. Scale -
1 cm. See Figure 15 for abbreviations.

On the dorsal surface a raised area begins medially and ante-
riorly but is missing farther in that direction, and it cannot be de-
termined if this separated the external nares.

A very thick tomial ridge is present on the ventral surface, and
a groove divided it from the shallower lingual ridge. Laterally, the
lingual ridge turns sharply backward. Many foramina penetrate
the triturating surface. In the midline a U-shaped groove, formed
between the lingual ridges, widens posteriorly.

The premaxilla protrudes anteriorly from the face in a manner
similar to the same bones in Podocnemis expansa.

218 FIELDIANA: GEOLOGY, VOLUME 16

MAXILLA
B. cooki

Earlier authors (Leidy, 1865; Hay, 1908) described the maxilla
as including the palatal pits. This study has revealed the fact that
the pits are formed by the jugals.

The maxilla (Figs. 13-16) may be divided into two main areas: a
curved anterior portion, somewhat vertical and exposed externally
on the cheek and a more horizontal portion sutured to the bones of
the palate and lying internal to the face. The external part meets
the premaxilla anteromedially and forms the lateral border of the
apertura narium externa. From the apertura to the orbit, there is an
irregular suture with the prefrontal. The maxilla forms a small
part of the orbital margin ventrally, meets the jugal dorso-laterally,
and ends in a broken edge posteriorly.

Anteriorly, the vertical portion of the maxilla is sutured to the
premaxilla and forms the floor of the fossa nasalis lateral to the
premaxilla and vomer. Posterior to the premaxilla the maxilla has
a short antero-posterior union with the vomer. At the junction with
the vomer, the palatino-maxillary suture turns laterad in the direc-
tion of the jugal pit. The maxilla meets the jugal just medial to the
pit and has a semi-circular course anteriorly around the pit.

The maxilla bears only a small portion of the labial ridge. The
ventral surface lateral to the labial ridge and the anterior edge of
the tomial ridge bear many nutritive foramina. The surface between
the labial and tomial ridges curves sharply into the deep jugal pit.

Opening onto the internal surface of the maxilla at the lateral
edge of the fossa nasalis lies the foramen alveolare superius.

PALATINE

B. cooki

Neither palatine is complete, but from surrounding bones the
limits may be determined. The palatine (Figs. 15, 16) has two
vertical surfaces which join laterally to form one sheet. Between
these surfaces medially lies the apertura narium interna. In front of
the apertura there is a short suture with the vomer; behind the aper-
tura the palatines meet each other for most of their length. The
palatine meets the maxilla anteriorly, the jugal laterally, and ptery-
goid posteriorly. In tranverse section the medial portion is C-shaped
around the apertura. The dorsal edge of the apertura is more pos-
terior than the ventral edge.

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 219

The labial ridge runs diagonally, antero-medially to postero-
laterally, across the palatine. The triturating surface has many
foramina and slopes up into the jugal pit. The medial surface has
no foramina and bends around onto the apertura.

On the dorsal surface a short posterior process rises to meet the
anterior part of the processus inferior parietalis. Between this ar-
ticulation and the dorsal edge of the apertura lies a small foramen
which does not seem to penetrate the ventral surface.

The palatine forms most of the posterior part of the orbital floor.

The foramen palatinum posterius penetrates the suture between
the palatine and the pterygoid about midway along the suture's
width, as in the living pelomedusid genera.

The overall relations of the palatine in Bothremys are quite simi-
lar to the condition in recent pelomedusids except for the presence of
a vomer in Bothremys. An internarial septum may be formed in the
recent genera by an anterior extension of the palatines in the midline.

VOMER
B. cooki

The vomer (Figs. 15, 16), a thin, medial element, extends from
the posterior junction of the premaxillae to the palatines and sep-
arates the aperturae narium externae. Its anterior end is broken so
the exact relations with the premaxillae are unknown, and it is pos-
sible that the vomer formed part of the margin of the foramen prae-
palatinum. Laterally, anterior to the aperturae the maxillae meet
the vomer in a fore-and-aft suture. The anterior half of the vomer
widens posteriorly to the aperturae where it narrows between the
two openings. The posterior end expands to meet the palatines,
but does not become as wide as the anterior half.

The vomer is higher than wide in cross-section and has a shallow
groove, the sulcus vomeri, running along its dorsal surface from front
to back.

The posterior half of the vomer rises dorsally above the level
of the median trough in the palate.

The vomer is absent in living pelomedusids except Podocnemis
vogli (Williams, 1954b, pp. 289, 291), but it is present in most chelyids
and amphichelydians. The vomer of Bothremys is more like that
bone in amphichelydians (particularly Chisternon hebraicum, AMNH
5961, from figures by Hay, 1908) than in Podocnemis vogli. The
importance of this distinction is debatable.

220 FIELDIANA: GEOLOGY, VOLUME 16

PTERYGOID

B. cooki

The pterygoids (Figs. 15, 19) are broken posteriorly and later-
ally; their limits are unknown in this region in B. cooki. These ele-
ments meet sagittally for about half their preserved length and have
a transverse suture with the palatines anteriorly. The small foramen
palatinum posterius penetrates the palate at the palatine-pterygoid
suture as in modern pelomedusids.

The processus pterygoideus externus meets the jugal anteriorly
beneath a thin covering of palatine now broken off and extends pos-
teriorly and laterally to form a process which, although poorly pre-
served, is comparable to that seen in the living pelomedusid genera.
The outer scroll-like portion is not preserved but the thicker, more
medial part is preserved and is essentially like that in Podocnemis.

Ventrally, the triangular basisphenoid lies between the posterior
halves of the pterygoid. A shallow groove parallels the basisphenoid-
pterygoid suture lateral to that suture on the ventral surface of the
pterygoid.

Dorsally, the pterygoids underlie the rostrum basisphenoidale an-
teriorly and extend posteriorly around the main body of the basisphe-
noid. The anterior limits are the same as on the ventral surface
but the posterior suture cannot be determined. The crista ptery-
goidea is fairly low but rises posteriorly. The foramen nervi trigemini
is much lower than in the living forms and cuts into the crista ptery-
goidea down to the level of the sulcus cavernosus. In Podocnemis
the foramen is rather high, while in Pelusios and Pelomedusa it is
lower but not at the level of the sulcus cavernosus. The sulcus itself
is quite wide, comparable in size to that seen in Podocnemis ma-
dagascariensis. A posterior pair of foramina, in the position of the
foramen pro ramo nervi vidiani is present in B. cooki but could not
be probed, so the actual identification is dubious. Another foramen
penetrates the medial edge of the right pterygoid anterior to the
above foramina; its purpose is unknown.

The enlarged "carotid canal" (Siebenrock, 1897; Williams, 1954a)
characteristic of the genus Podocnemis is absent in Bothremys.

B. barberi

The Selma specimen, although in much worse condition, does
preserve the postero-medial limits of the pterygoids (Fig. 19) . The
area is crushed, weathered, and has poorly defined sutures.

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 221

The pterygoid of living pelomedusids extends posteriorly less
than half the length of the basisphenoid. The pterygoid of the type
specimen extends at least as far back as the end of the basisphenoid,
and the Selma specimen shows the pterygoids reaching the preserved
limits of the basioccipital (which includes most of the condyle).

The enlarged carotid canal of Podocnemis (Fig. 19) is absent,
and there is an indication of a quadrate connection to the basisphe-
noid and possibly to the basioccipital.

The pterygoid of living pelomedusids does not floor any part of
the fenestra postotica, but, due to the posterior extension of this bone
in Bothremys, it may have underlain part of this cavity. However,
the area is badly damaged or missing, and this cannot be verified.

The foramen posterius canalis carotid interne as well as the rela-
tionship of the pterygoid to the exoccipital and opisthotic are not
determinable.

BASISPHENOID
B. cooki

The posterior and postero-lateral limits of the basisphenoid
(Fig. 19) are unknown. It has a triangular shape ventrally with
the apex cut by a short transverse suture. The preserved posterior
edge may be the actual one, but being badly damaged this cannot be
determined. Its anterior and lateral boundaries are sutured to the
pterygoids.

Dorsally, the rostrum basisphenoidale extends anteriorly over the
sagittal suture of the pterygoid to the posterior limits of the palatines.
The sella turcica is very deep and somewhat overhung by the dorsum
sellae much as in the three living genera (except Podocnemis mada-
gascariensis) . In the lateral wall of the sella turcica lies the large
paired foramen anterius canalis carotid interni. The processus cli-
noideus is absent in Bothremys but present in the specimens of liv-
ing pelomedusid genera examined by us (although it is reduced in
Pelusios gabonensis) . Posterior to the dorsum sellae the basisphenoid
is a concave plate.

The area of the cavum labyrinthicum and hiatus acusticus is bad-
ly damaged and undeterminable. No other foramina are preserved.

B. barber -i

Although most of the basisphenoid is preserved in this specimen,
the preservation is so poor that little more can be added.

222 FIELDIANA: GEOLOGY, VOLUME 16

Fig. 17. Bothremys barberi, spp. B (FMNH PR 247), Selma Formation, Ala-
bama, ventral view of interior of cranial cavity and right otic region. Left side of
scale is in millimeters. See Figure 15 for abbreviations.

The acustic hiatus is at the posterior end of the basisphenoid on
the basioccipital suture. The ventral suture with the basioccipital
can also be seen. It curves anteriorly and is similar to that in Po-
docnemis. The postero-lateral limits cannot definitely be deter-
mined, but there seems to be contact with the quadrate.

The rest of the specimen agrees with the type except for the
shallower sella turcica which may be due to breakage.

BASIOCCIPITAL

B. barberi

Dorsally the basioccipital is a small triangular element with the
apex pointing posteriorly. It has a transverse suture anteriorly

Fig. 18. Bothremys barberi, spp. B (FMNH PR 247), Selma Formation, Ala-
bama, skull table fragment and right otic process. A, dorsal view. B, lateral
view. BS - basisphenoid, CT - cavum tympani, F - frontal, ICA - incisura cclu-
mellae auris, OP - opisthotic, P - parietal, PR - prootic, PRF - prefrontal, Q -
quadrate, SO - supraoccipital, SQ - squamosal. Basisphenoid has been crushed
upward. Scale - 1 cm.

I

223

224 FIELDIANA: GEOLOGY, VOLUME 16

with the basisphenoid and postero-lateral sutures with the two exoc-
cipitals. It is excluded from the foramen magnum by the exoc-
cipitals.

Ventrally there is an anterior suture with the basisphenoid
which curves anteriorly as in the living forms. The precondylar
fossa is delimited by this suture and is wider than long. The pos-
terior surface and condylar area are badly eroded and sutures can-
not be determined. Pelusios and Pelomedusa lack this fossa which
is best developed in Podocnemis, particularly P. expansa. This
area in Bothremys is similar to P. expansa and differs principally in
being shorter. The quadrate may possibly articulate with the lat-
eral extension of the basioccipital, but the sutures are badly obscured,
and this seems unlikely.

The tubercula basioccipitalia are large and prominent in Podocne-
mis but absent in Pelusios and Pelomedusa. This area in Bothremys is
difficult to interpret, but the tubercula seem to be lacking entirely.

SUPRAOCCIPITAL

B. cooki

The supraoccipital (Figs. 13, 14, 16) lacks most of its posterior
portion (including the crista supraoccipitalis) , and its anterior rela-
tions are obscured by damage from pyrite decomposition.

On the ventral surface the supraoccipital has a diagonal suture
with the parietal trending posteriorly from the midline. This suture
turns anteriorly on the medial edge of the foramen nervi trigemini.
The actual edge of the foramen, however, is bordered by the parietal
dorsally, the prootic posteriorly, and the pterygoid ventrally. The
relations of the prootic to the supraoccipital in this area are not clear.
The supraoccipital has long interdigitations along its anterior edge
with the parietal and prootic, which makes suture determination
difficult in eroded areas.

The recessus labyrinthicus supraoccipitalis is preserved on the
left side. The crus commune is contained here as well as the dorsal
portions of the canalis semicircularis anterior and the canalis semi-
circularis posterior. The upper edge of the hiatus acusticus is badly
broken.

Fig. 19. A, B, Bothremys cooki (type skull), basisphenoid and pterygoids.
A, ventral. B, dorsal. C, Podocnemis unifilis (AMNH 58200), recent, South
America, ventral view of right posterior portion of skull. D, Bothremys barberi
(FMNH PR 247), same region as C. Scale - 1 cm. Abbreviations on page 236.

225

226 FIELDIANA: GEOLOGY, VOLUME 16

Externally only the most anterior part of the supraoccipital is
preserved. It articulates with the parietal dorso-anteriorly. Its
other sutures cannot be determined adequately.

B. barberi

The supraoccipital (Figs. 17, 18), agrees with B. cooki. The
dorsal sutures, however, are slightly better preserved. The parietal-
supraoccipital suture is directed more ventrally than in B. cooki.
The ventral margin has a crushed contact with the opisthotic and
may have met the posterior edge of the prootic.

OPISTHOTIC
B. barberi

The opisthotic meets the prootic anteriorly in a long suture,
the quadrate dorso-laterally, the squamosal postero-laterally, and
the supraoccipital medially. The area of this last suture is badly
broken and pushed up onto the supraoccipital. Relations with the
exoccipital are unknown.

Ventrally (internally), the opisthotic (Fig. 17) forms the pos-
terior wall of the labyrinth. A broken processus interfenestralis may
be seen, but further details cannot.

PROOTIC

B. barberi

The right prootic (Fig. 18) is crushed and poorly preserved; the
left one is absent. Dorsally, the prootic has a short suture with the
parietal antero-medially, the quadrate laterally, the opisthotic pos-
teriorly, and (?) the supraoccipital postero-medially. Ventrally
(Fig. 17) the semi-circular canals are preserved. The prootic forms
the anterior wall of the labyrinth, but details such as the fossa
acustico-facialis cannot be seen. No additional information can be
obtained from B. cooki.

QUADRATE
B. barberi

Only the right quadrate (Figs. 17, 18, 19) is preserved. It has
been crushed and probably distorted, but the more important rela-
tionships are determinable. The anterior, posterior, and lateral

GAFFNEY AND ZANGERL: THE BENUS BOTH REM YS 227

edges are eroded, and the bone is split through the middle revealing
some of the internal structure.

The quadrate is relatively more extensive laterally than in living
pelomedusids. The anterior edge has been eroded back to the for-
amen stapedio-temporale. The medial border is sutured to the pro-
otic anteriorly and the opisthotic posteriorly. The posterior margin
is badly broken, but there seems to be a small portion of the squa-
mosal preserved just lateral to the posterior termination of the
opisthotic-squamosal suture. The whole lateral edge is represented
by an eroded surface.

Much of the ventro-medial structure is obscured by crushing
in which the basicranium has been pushed dorsally and laterally
into the otic extension. The existence of a processus trochlearis is
not known. The form of the cavum acustico-jugulare is not deter-
minable.

At the postero-ventral extension of the quadrate lies the proc-
essus articularis (Fig. 19). It has a greater ventral extent than in
living pelomedusids. Furthermore, one of the chief differences be-
tween Bothremys and the living genera is the posterior position of
the processus articularis in the extinct genus. Modern pelomedusids
have the articulation quite forward of the occipital condyle, in fact,
well anterior to the basioccipital. In Bothremys, however, the proc-
essus articularis is in the same vertical plane with the occipital con-
dyle, a condition similar to amphichelydians, and, therefore, prob-
ably primitive. Much of the actual articulating surface, the con-
dylus mandibular is, is eroded, but its shape can be determined. It
had the form of an elongate oval and did not seem to be sharply
concave as in the recent forms, but slightly concave or flat.

Just medial to the processus articularis is the postero-lateral
extension of the pterygoid which forms a process buttressing the
quadrate.

The edges of the cavum tympani (Fig. 18) are eroded and broken
so its complete extent is not known. As preserved, it is an oval
depression with a foramen near its anterior border. Internally, it
can be seen that this foramen is the incisura columellae auris, in
this case, actually a canalis because of the posterior closure of the
quadrate. It is far removed from the posterior edge of the cavum
and is more ossified than in any other pleurodire. Slightly posterior
to the stapedial opening is a small indentation, possibly a post-otic
antrum, but the area is poorly preserved. The cavum tympani also
contains a large trough that extends postero-ventrally from the in-

228 FIELDIANA: GEOLOGY, VOLUME 16

cisura and medially behind the quadrate to the vicinity of the fenestra
postotica.

A fracture through the middle of the otic region has exposed
some internal details (Fig. 17). The incisura columellae auris ex-
tends from the lateral surface of the quadrate medially to a canal,

ANT. M \ POST.

Fig. 20. Bothremys cooki (type skull), ?Hornerstown Formation, New Jersey,
cross-section through a cast of the right palatal pit. J - jugal, M - maxilla.

which probably represents the cavum acustico-jugulare. The region
is not sufficiently well preserved to compare with other forms.

ENDOCRANIAL CAST

Endocranial casts (Fig. 21) of latex were made from both spec-
imens. Although it is generally true that reptilian endocasts do
not reflect much detail of the brain itself, they are informative to
some extent.

The endocranium of B. cooki is best preserved and its cast may
be compared with a similar cast of Podocnemis Madagascar iensis.
As Edinger (1929) has shown, the first principal flexure generally
delimits the cerebral area from the olfactory region. The sulcus
olfactorius which extends anteriorly from this flexure is wide in Both-
remys and narrow in P. madagascariensis. The sides of this trough
are also more extensive in the extinct genus. The cerebral expansions
are similar in both but wider in Bothremys. The unossified anterior
portion of the supraoccipital has left a "rider" on top of the cerebral
area (see Zangerl, 1960, p. 290). Just behind the hemisphere region
and below the "rider" is a low dorso- ventral groove leading to the
foramen nervi trigemini.

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 229

Fig. 21. Endocranial casts, A, C, E, dorsal view; B, D, lateral views. A, B,
Podocnemis madagascariensis (AMNH Living Reptiles 63579). C, D, Bothremys
cooki (type skull). E, Bothremys barberi (FMNH PR 247). CH - region of cerebral
hemispheres, CR - cartilaginous "rider" fill (see text), FM - position of foramen
magnum, FNT - foramen nervi trigemini, SF - sulcus olfactorius. Scale - 1 cm.

The endocranium of Bothremys barberi has less preserved along
the sides, but comparisons can be made. It differs from B. cooki
in having much less pronounced cerebral expansions and a relatively
smaller sulcus olfactorius. The differences are greater between B.
cooki and B. barberi than between B. cooki and P. madagascariensis.
This apparent anomaly may be explained by the absolute size dif-
ference in the extinct skulls. Reptilian brains tend to be relatively
smaller in larger animals. B. barberi is twice the size of B. cooki,
but the brains were probably only slightly different in size (if recent

230 FIELDIANA: GEOLOGY, VOLUME 16

reptiles are a correct analogy). Therefore, although the brains
were probably very similar, the cavities housing them are not.

MANDIBLE

B. cooki

The posterior parts of the jaw (Fig. 22), including articular sur-
faces, have been broken off. The specimen has suffered from pyrite
decomposition and none of the sutures are distinguishable.

The most obvious features are the large pits that occupy most
of the jaw. These excavations open anteriorly and extend poste-
riorly as conical cavities well beneath the coronoid processes. The
whole anterior half of the jaw is concerned with forming these pits.

There is no distinct ridge anteriorly; the surface is essentially
horizontal. Posteriorly the labial margin curves upward and forms
the lateral side of the trumpet-shaped pit. The tomial ridge begins
near the symphysis and extends posteriorly and dorsally, forming the
medial edge of the pit.

The symphysis is tightly fused. It is much thicker posteriorly
than anteriorly, forming a triangular cross-section.

The coronoid process is quite high and well developed. It is
rounded dorsally and somewhat longer than wide.

The inner face of each ramus (Fig. 22B) exposes the fossa meckelii
and the sulcus cartilaginis meckelii. Dorsally the fossa is poorly
preserved and plastered up, but more ventrally, the sulcus carti-
laginis meckelii can be seen extending anteriorly. The foramen for
the inferior maxillary nerve, or the ostium inferius inframaxillaris
of Ogushi (1911), is present just above the ventral border of the
fossa.

The jaw is 54 mm. wide at the coronoid processes while the Selma
specimen is 105 mm. wide at that point.

B. barber i

This specimen (Fig. 22D, E) agrees for the most part with B.
cooki, but the posterior parts are preserved and some (but not all)
of the sutures may be seen. Most of the bone is weathered and
eroded.

Fig. 22. A, B, C, Bothremys cooki (type). D, E, Bothremys barberi (FMNH
PR 247). A, lateral - dotted lines indicate extent of pit. B, medial. C, dorsal.
D, ventral. E, dorsal. Scale - 1 cm. Abbreviations on page 236.

231

232 FIELDIANA: GEOLOGY, VOLUME 16

The dentaries agree with B. cooki, but the symphysis lacks the
thickened posterior edge. The pits are relatively smaller (only
slightly) and the bone is thicker in the Selma jaw. This may be
due to the larger size of this specimen (about twice the size of B.
cooki.)

What appears to be a suture defining the coronoid may be fol-
lowed on the left ramus and partially on the right. Dorsally it is
behind the coronoid process, about one-third of the distance be-
tween that and the articulation. It proceeds down the lateral side
slanting anteriorly until it is about halfway down the jaw. There
it seems to split into two sutures, one going ventrally, the other
curving anteriorly. The anterior one (the coronoid) goes into the
pit and bisects the structure at least partially. It continues out
the medial edge of the pit, slanting dorsally. On the medial surface
of the jaw this suture makes a wide ventral curve, bordering the
dorsal part of the fossa meckelii, then curving dorsad to just behind
the coronoid process. The coronoid, then forms the upper half of
the conical pit. This contrasts with the coronoid of living pelo-
medusids in which it extends only slightly, if at all, onto the trit-
urating surface.

The dentary has a long ventral process that extends posteriorly
as in living pelomedusids.

The medial border of the fossa meckelii is broken. It is the same
as in B. cooki except that here the dorsal limits are exposed. The
shape is quite similar as in the living genera. The antero-posterior
length of the fossa is greater than in Chisternon hebraicum (AMNH
5904). The position of the ostium inferius canalis inframaxillaris
is very similar in Bothremys and Chisternon as well as in the living
pelomedusids. As in B. cooki, the sulcus cartilaginis meckelii ex-
tends to just behind the symphysis.

The area articularis mandibularis forms the posterior border of
the fossa meckelii. The overall shape of the area articularis is most
similar to Podocnemis madagascariensis. The surface is slightly
convex and nearly circular in outline. Most other pelomedusids
have strongly convex surfaces. A retroarticular process, slightly
longer than the one in P. madagascariensis extends posteriorly from
the area articularis mandibularis.

A small portion of the right squamosal as well as parts of the
exoccipitals are preserved in B. barberi, but they are too fragmentary
and crushed to describe.

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 233

RELATIONSHIPS

The skull of Bothremys possesses the following characteristics
which demonstrate that the genus is a member of the Suborder Pleu-
rodira as currently defined (Romer, 1956) :

1) Frontal enters orbital margin (also in Chelonia mydas and
Corsochelys haliniches, Zangerl, 1960; p. 287; see also Simpson,
1938);

2) prefrontal lacks descending process to vomer;

3) quadrate closed behind stapes;

4) articular surface on mandible slightly convex.

The following shell characters support this assignment:

1) fusion of ilia to carapace and fusion of pubes and ischia to
plastron ;

2) presence of lateral mesoplastra.

The two families of pleurodires universally recognized are the
Pelomedusidae and the Chelyidae (Zangerl, 1948; Romer, 1956).
The skull of Bothremys has the following pelomedusid characters
which are lacking in chelyids:

1) prefrontals large and in sagittal contact for all their length;

2) nasals absent;

3) lower jaw with broad triturating surfaces.

The presence in Bothremys of a vomer, however, is a chelyid char-
acter. A vomer also exists in Podocnemis vogli, an otherwise ac-
ceptable pelomedusid. This fact together with the inference that
animals not far removed from the chelyid-pelomedusid split might
all retain vomeres, tends to lessen the strength of this character. The
lateral mesoplastra of the genus Bothremys are also diagnostic of
pelomedusids.

Relationships beneath the family level are more difficult to de-
termine. Zangerl (1948, pp. 39-52) has shown that the shell of
"Podocnemis alabamae" (Bothremys barberi) is most similar to the
shell of the living South American species of Podocnemis, and he
suggested that "P. alabamae" was a good morphologic ancestor to
them. The skull, on the other hand, was considered to be quite
distinct from the living genera as early as 1891 when Bauer (1891,
p. 423) erected the family Bothremydidae based on this genus. Put-
ting aside for the present the possibility of its being placed in a sep-
arate family, the jugal specializations of Bothremys (see descriptive

234 FIELDIANA: GEOLOGY, VOLUME 16

section) remove it from the direct ancestry of any pleurodires in
which the skull is known.

Some positive statements, however, can be made about more
general relationships. The posterior position of the quadrate and
the well-developed vomer are found together only in Bothremys.
This suggests that Bothremys may be closer to the ancestral pelo-
medusid condition than are the other pelomedusid genera. This
is a very shaky supposition, but it may be strengthened by exam-
ining the pelomedusids known by skulls. Podocnemis (Siebenrock,
1897) and Dacquemys (Williams, 1954a) have an enlarged "carotid
canal," a feature which is "advanced" for pleurodires (see Williams,
1954a). Stereogenys (Andrews, 1906, p. 297; Dacqu£, 1912; Williams,
1954a) has an enlarged carotid opening as well as a well-developed
secondary palate. The areas in question are not preserved in Shwe-
boemys (Swinton, 1939). Pelusios and Pelomedusa, although hav-
ing "normal" carotid canals, also have the quadrate far anterior
to the occiput (as in the preceding genera : Podocnemis, Pelomedusa,
Pelusios, Stereogenys, and Dacquemys), a condition not found in
amphichelydians. Carter emys (Williams, 1953; Carter, 1852) has
not been described in enough detail to determine the features being
discussed here. Therefore, although Bothremys is highly specialized
in the structure of the cheek and triturating surfaces, it is more
primitive than known pelomedusids in its other characters:

1) presence of a vomer;

2) absence of enlarged carotid canals;

3) posterior position of quadrate.

The shell morphology of Bothremys may be reconciled with this
information from the skull. The similarity of the shell of the South
American Podocnemis with Bothremys might be due to convergence,
but it more probably indicates that this is a primitive pattern among
pelomedusids and was retained by Podocnemis (it is also found in
Roxochelys, see Price, 1953, an upper Cretaceous pelomedusid from
Brazil) .

In conclusion, Bothremys has features, such as the presence of
the jugals on the triturating surfaces (not found in any other known
chelonian), that indicate considerable divergence from what would
be considered a "generalized condition." However, other evidence
from the skull and shell suggests that Bothremys represented a lower
level of organization within the pelomedusids than has been recog-
nized in other pelomedusid genera.

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 235

THE ADAPTATIONS OF BOTHREMYS

The function of the large palatal and mandibular pits in Both-
remys has been commented on by earlier authors. Leidy (1865,
p. Ill) wrote, "The function of the latter (the pit) ... is difficult
to comprehend. It does not appear to be an alveolus for a tooth; but
probably it may have accommodated a corneous tooth-like process
springing from a corresponding hollow of the lower jaw." Baur
(1891, pp. 423-424) suggested the presence of "a large tusk" in the
pits. Hay (1908, p. 104) believed that the tooth idea was improb-
able and made another suggestion: "The whole construction of the
skull of Bothremys indicates that it was accustomed to crush hard
objects as food. Probably these objects were of such a nature that
economy of force demanded that they should be brought to a par-
ticular spot on the jaw for crushing. To provide for the rapid repro-
duction of the horn beneath these areas for crushing, these pits became
developt in a way analogous to the human 'nailbed.' "

Before discussing the merits of these ideas, the morphology of
the structures involved should be summarized. There are two pits
in the palate, principally formed by the jugals, but all of the sur-
rounding bones including the cheek are modified to form the pits.
Each pit is thickest (see cross-section) about midway in the cone.
The triturating surfaces occupy most of the palate and have an
overall form of two funnels opening ventrally. The lower jaws
have matching structures which, however, do not open directly
upward into the jugal pits, but open dorso-anteriorly.

Hay was almost certainly right in thinking that Bothremys was
toothless. The "corneous tooth-like process" of Leidy has a some-
what greater possibility; however, the horny covering of recent
turtles rarely forms a structure that is not present in the underlying
bone. Thicknesses vary, but whole structures, like a large tusk,
are not known to occur independent of an osseous core. Therefore,
it seems likely that the whole triturating surface was covered by
horny material, thickest in the tips of the pits but not forming
tusk-like structures.

This inference is most like the suggestion of Hay previously
quoted. However, Hay also thought that the pits were areas of
rapid horn production and were not expressed on the external sur-
face. It would be difficult to disprove this, but it is more probable
that the pits were expressed as depressions in the external horny
covering as in living turtles.

236

FIELDIANA: GEOLOGY, VOLUME 16

If this latter suggestion were correct, the pits might function to
crack hard objects, slightly larger than the pits, which otherwise
would be difficult to grasp or hold onto by means of a "normal"
triturating surface. Certain ovoid molluscs (gastropods?) might
comprise the food of Bothremys. There is a disadvantage to this
hypothesis since it seems to mean that a particular individual of
Bothremys would be limited to objects of a certain size range, be-
cause larger or smaller ones could not fit the pits. The crushing
function of the pits is substantiated by other structural evidence.
High and well-developed coronoid processes, as in Bothremys (Fig.
22A,B), are characteristic of turtles with crushing habits (see Stej-
neger, 1944, plate 30, for an example in Amyda). The posterior
position of the articulation and the position of the pits immediately
adjacent to the coronoid processes result in an efficient crushing
mechanism with the adductor musculature attached close to the
object to be crushed, but far from the fulcrum.

ABBREVIATIONS

AM - area articularis mandibularis

ANE - apertura narium externa

BO - basioccipital

BS - basisphenoid

CH - region of cerebral hemispheres

CL - cavum labyrinthicum

CM - condylus mandibularis

CO - condylus occcipitalis

COST - costal

CP - crista pterygoidea

CR - cartilaginous ("rider") fill of
supraoccipital

CT - cavum tympani

DS - dorsum sellae

E - epiplastron

ENT - entoplastron

EO - exoccipital

F - frontal

FAC - foramen anterius canalis caro-
tid interni

FCI - foramen posterius canalis carot-
id interni

FM - region of foramen magnum

FNT - foramen nervi trigemini

FPC - foramen posterius canalis carot-
id interni

FPP - foramen palatinum posterius

FPR - foramen praepalatinum

FRV - foramen pro ramo nervi vidiani

FS - fossa mecklii

HA - hiatus acusticus

HYO - hyoplastron

HYPO - hypoplastron

ICA - incisura columellae auris

J - jugal

LP - lateral process of parietal

M - maxilla

ME - mesoplastron

N - neural

NUC - nuchal

OH - ostium inferius canalis infra-

maxillaris (see Ogushi, 1911)
OP - opisthotic
P - parietal

PC - processus coronoideus
PE - peripheral
PG - pygal

PIP - processus inferior parietalis
PL - palatine
PM - premaxilla
PO - postorbital
PPE - processus pterygoideus exter-

nus
PR - prootic
PRF - prefrontal
PT - pterygoid
Q - quadrate

RBS - rostrum basisphenoidale
SCM - sulcus cartilaginis meckelii
SET - sella turcica
SF - sulcus olfactorius
SOC - supraoccipital
SPG - suprapygal
SQ - squamosal
SU - sulcus cavernosus
V - vomer
X - xiphiplastron

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 237

INSTITUTIONS:

AMNH - American Museum of Natural History
ANSP - Academy of Natural Sciences of Philadelphia
FMNH - Field Museum of Natural History

REFERENCES

Andrews, C. W.

1906. A descriptive catalogue of the Tertiary vertebrata of the Fayum, Egypt.
Trustees of the British Museum, London, 324 pp.

Baird, D.

1964. A fossil sea-turtle from New Jersey. N. J. State Mus. Inves., No. 1,
pp. 3-26.
Baur, G.

1891. Notes on some little-known American fossil tortoises. Proc. Acad. Nat.

Sci. Phil., pp. 411-430.
1893. Notes on the classification and taxonomy of the Testudinata. Proc.
Amer. Phil. Soc, 31, pp. 210-225.

Carter, H. J.

1852. Geology of the island of Bombay. Jour. Bombay Branch. Roy. Asiatic
Soc, 21, pp. 161-215.

Cobban, W. A. and Reeside, J. B.

1952. Correlation of the Cretaceous formations of the western interior of the
United States. Bull. Geol. Soc. Amer., 63, pp. 1011-1044

Collins, R. L. and Lynn, W. G.

1936. Fossil turtles from Maryland. Proc. Amer. Phil. Soc, 76, (2), pp. 151-
174.

Cooke, C. W., et al

1943. Correlation of the Cenozoic formations of the Atlantic and Gulf coastal
plain and the Caribbean region. Bull. Geo. Soc Amer., 54, pp. 1713-1723.

Dacque, E.

1912. Die fossilen Schildkroten Aegyptens. Geol. Palaeont. Abhandl., 14,
pp. 275-337.

Dorf, E., ed.

1957. Guidebook for field trips, Atlantic City meeting. Geol. Soc Amer.
pp. 1-280.

E dinger, T.

1929. Die fossilen Gehirne. Zschr. ges. Anat., Abt. Ill, 28, 249 pp.

Hay, O. P.

1908. The fossil turtles of North America. Carnegie Inst., Publ. no. 75, 568 pp.

Kesteven, H. L.

1910. The anatomy of the head of the green turtle Chelone midas Latr. Part
I. The skull. Jour. Proc. Roy. Soc New South Wales, 44: pp. 368-400.

Leidy, J.

1865. Memoir on the extinct reptiles of the Cretaceous formation of the United
States. Smithsonian Contributions to Knowledge, 14, art. 6, pp. 1-135.

238 FIELDIANA: GEOLOGY, VOLUME 16

Miller, H. W.

1955. A check-list of the Cretaceous and Tertiary vertebrates of New Jersey.
Jour. Paleontol., 29 (5), pp. 903-914.

1956. Correlation of Paleocene and Eocene formations and the Cretaceous-
Paleocene boundary in New Jersey. Bull. Amer. Assoc. Petrol. Geol., 40
(4), pp. 722-736.

Nick, L.

1913. Das Kopfskelett von Dermochelys coriacea L. Zool. Jahrb. Abt. Anat.
Onotog. Tiere, 33, pp. 1-238.

Ogushi, K.

1911. Anatomische Studien an der japanischen dreikralligen Lippenschildkrote
(Trionyx japonicus). 1. Mitteilung. Morph. Jahrb., 43, pp. 1-106.

Olsson, R. K.

1963. Latest Cretaceous and earliest Tertiary stratigraphy of New Jersey
coastal plain. Bull. Amer. Assoc. Petrol. Geol., 47 (4), pp. 643-665.

Parsons, T. S. and Williams, E. E.

1961. Two Jurassic turtle skulls: a morphological study. Bull. Mus. Comp.
Zool., 125 (3), pp. 43-107.

Price, L. I.

1953. Os quelonios da formacao Bauru, Cretaceo terrestre do Brasil meridional.
Dept. Nac. da Prod. Min., Div. de Geo. e Min., Boletim 147, pp. 7-34.

PvOMER, A. S.

1956. Osteology of the reptiles. Univ. of Chicago, 772 pp.

Schmidt, K. P.

1940. A new turtle of the genus Podocnemis from the Cretaceous of Kansas.
Fieldiana: Geol., 8 (1), pp. 1-12.

SlEBENROCK, F.

1897. Das Kopfskelett der Schildkroten. Sitzungsb. K. Akad. Wiss. Vienna
(Math.-Nat. Kl.) 106, Abt. I, pp. 245-328.

Stejneger, L.

1944. Notes on the American soft-shell turtles with special reference to Amyda
agassizii. Bull. Mus. Comp. Zool., 94 (1), pp. 1-76.

Stephenson, L. W.

1942. Correlation of the outcropping Cretaceous formations of the Atlantic
and Gulf coastal plain and Trans-Pecos Texas. Bull. Geol. Soc. Amer., 53,
pp. 435-448.

Swinton, W. E.

1939. A new fossil fresh water tortoise from Burma. Records of the Geol.
Surv. of India, 74, pt. 4, pp. 548-551.

Williams, E. E.

1953. Fossils and the distribution of chelyid turtles. 1. "Hydraspis leithii"
in the Eocene of India is a pelomedusid. Breviora, no. 13, pp. 1-8.

1954a. New or redescribed pelomedusid skulls from the Tertiary of Africa and
Asia (Testudines, Pelomedusidae). Breviora, no. 35, pp. 1-8.

1954b. A key and description of the living species of the genus Podocnemis.
Bull. Mus. Comp. Zool., Ill (8), pp. 279-295.

GAFFNEY AND ZANGERL: THE GENUS BOTHREMYS 239

Zangerl, R.

1948. The vertebrate fauna of the Selma Formation of Alabama. Part I —
Introduction. Part II — The pleurodiran turtles. Fieldiana: Geol. Mem., 3
(1 and 2), pp. 1-56.

1953. The vertebrate fauna of the Selma Formation of Alabama. Part III —
The turtles of the family Protostegidae. Part IV — The turtles of the family
Toxochelyidae. Fieldiana: Geol. Mem., 3 (3 and 4), pp. 57-277.

1960. The vertebrate fauna of the Selma Formation of Alabama. Part V — An
advanced cheloniid sea turtle. Fieldiana: Geol. Mem., 3 (5), pp. 281-312.

Zangerl, R. and Sloan, R.

1960. A New Specimen of Desmatochelys lowi, Williston. A primitive Chel-
oniid Sea Turtle from the Cretaceous of South Dakota. Fieldiana: Geol., 14
(2), pp. 7-40.